a b s t r a c ta r t i c l e i n f o
Article history:
Received 31 March 2014
Received in revised form 28 July 2014
Accepted 3 August 2014
Available online 10 August 2014
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
Hundreds of catastrophic floods resulting from man-made dam failures
have been recorded across the globe (e.g., Costa, 1985; Peng and
Zhang, 2012), including those in the Czech Republic (see Table 1). This
specific type of flood is characterised by extreme discharges that frequently
exceed the maximal flow rates of ‘classical’ hydrometeorological
floods by several times. Thus, this type of flood is also well
characterised by its great potential to cause significant material damage
and to claim lives. In addition, geomorphologic and landscape effects are
often considerable and identifiable after years (e.g., Jarrett and Costa,
1986; Costa and O'Connor, 1995; Bathurst and Ashiq, 1998).
This article aims to reconstruct the geomorphologic and landscape effects
of the most catastrophic flood from a man-made dam failure within
the territory of today's Czech Republic, namely, the Bílá Desná dam failure,
which occurred on 18 September 1916, only one year after the completion
of its construction. At the same time, the specific aim of this paper
is to evaluate and to discuss methodological frameworks for the use
of documentary data that are currently accepted by geomorphologists
(Glade et al., 2001), mainly in respect of data types and their
analysis.
Despite the facts indicating that the volume of impounded water
was relatively low (~290,000 m3
) in comparison with recorded failures
of earthen dams worldwide and that people living downstream were
warned approximately 30 min before the dam failure, this event directly
claimed 62 lives and also caused significant material damage (Žák
et al., 2006). Most victims were registered in the village of Desná,
which is situated ~5 km downstream from the position of the former
reservoir.
Most studies of this event have predominantly focused on geotechnical
causes, mechanisms and socioeconomic effects of the catastrophe.
These studies include early works presented, e.g., by Smrček (1917),
Gebauer (1920), Gnendiger (1920), or Karpe (1935) and later works
presented by Mikolášek (1966), Žák (1996), or Žák et al. (2006). Information
concerning hydrological significance of the flood, geomorphologic
and landscape effects are found in these works and are generally
noncomprehensively presented in the form of a by-product of other
investigations (e.g., court notes). Therefore, the compilation of this
fragmented information from various data types and sources and the
production of a new aggregate context supported by newly presented
analysis and reconstructions will fill the gap in the current knowledge
of this event.
2. Study area and characteristics of the former Bílá Desná dam
The Bílá Desná River springs in Jizerské hory Mts. (northern
Bohemia, Czech Republic) in a peat-bog called ‘Na Kneipě’ at an elevation
of approximately 1000 m asl on the southern slopes of the Jizera
Table 1
Reported man-made dam failures in the Czech Republic.
Dam Construction Failure Cause (trigger) Escaped volume
[×1000 m3
]
Loss and damages References
Strážný 14th Century 1587 Not known Not known Not known http://www.turistika.cz/mista/
strazny-prehrada
Mladotický pond 14th Century 25.5.1872 Intense rainfall About 3000.00 Several fatalities, significant
material damages
http://www.nebreziny.cz/
Bílá Desná 1915 18.9.1916 Insufficient geotechnical
survey
290.00 62 victims, significant
material damages
Gnendiger (1920); Žák et al. (2006)
Zlatá Ktič 1796 2002 Intense rainfall 98.50 None http://www.sobenov.cz/mis/
Soběnov 1925 2002 Flood wave from Zlatá Ktiš
dam failure
73.20 None http://www.sobenov.cz/mis/
Blažnov pond – 28./29.3.2006 Intense rainfall/snowmelt 10.37 None Šobr et al. (2008)
Fig. 1. Study area of the Bílá Desná River and the position of the former dam.
136
hill (1122 m asl). The river, with a length of ~12 km and a catchment of
16.0 km2
, joins the Černá Desná in the village of Desná to form the
Desná River (Fig. 1), which is a left-sided tributary of the Kamenice
River belonging to the Labe River catchment. The Jizerské hory Mts.
are considered one of the rainiest parts of the Czech Republic. The average
annual precipitation in the highest parts exceeds 1200 mm/y
(ČHMÚ, 2013), and the maximal 24 h precipitation in the Czech
Republic was measured at Nová Louka, which is situated in the western
part of the Jizerské hory Mts. (345 mm; 28–29 July 1897), causing the
most catastrophic flash floods in mountain ranges at the northern borderland
of the country (Štekl et al., 2001).
The flood that followed the above-mentioned heavy rainfalls at the
end of July 1897 caused significant material damage in densely settled
valleys and initialised a discussion concerning the construction of a system
of dams to prevent (mitigate) potentially catastrophic effects of future
floods (Žák, 1996). This effort resulted in the construction of several
dams during the first two decades of the twentieth century in the
Jizerské hory Mts. and in its foothills, including the masonry arch dam
Bedřichov at Černá Nisa, which was built between 1902 and 1909; the
masonry arch dam Mšeno at Jablonec nad Nisou, which was built between
1906 and 1908; the masonry arch dam Fojtka, which was built
at Mníšek between 1904 and 1906; the masonry arch dam Harcov,
which was built at Liberec between 1902 and 1906; the embankment
earthen dam Souš at Černá Desná, which was built between 1911 and
1915; and the embankment earthen dam Desná at Bílá Desná, which
was built between 1911 and 1915 (Fig. 2).
2.1. Characteristics of the former Bílá Desná dam
The Desná dam was built in the upper part of the Bílá Desná River,
~5 km far from its spring, at an altitude of 806 m asl. The dam was
designed to be a 14.16-m-high embankment earthen dam with one outlet
controlled from the spool tower. The reservoir was designed to retain
400,000 m3
of water. Excess water was distributed through the tunnel
to the Souš reservoir, which was situated in the adjacent valley, and
spillways were available. The basic characteristics of the former Desná
dam and reservoir are listed in Table 2.
3. Data sources and methods
3.1. Historical data sources
Documentary data represent an important source of information for
various types of historical natural hazards where the natural proxy indicators
are missing or were erased by subsequent human activity.
Although the spectrum of documentary data is quite broad (Ingram
et al., 1981; Hooke and Kain, 1982), only a few types are used in
geomorphology, which can be ascribed to the prevailing focus on
the creation of historical time series of natural hazards (cf. Glade
et al., 2001). Considering the specific aim of this paper, which is to
reconstruct a single historical natural disaster, we employ different
types of data with respect to their time dimension and purpose of
origin, with special attention devoted to stationary and quasistationary
sources (Raška et al., 2014). Furthermore, these data
were subject to critical analyses before the information concerning
the studied flood was cited. Table 3 shows the summary of the data
that were analysed in this study, together with their types and
sources. Most of these data were acquired by detailed study in the
regional archive of the Jablonec nad Nisou in northern Czechia;
however, national and regional archives, libraries, personal collections,
and web repositories were also searched.
Fig. 2. Unique archival map showing the position of the Bílá Desná dam (1) in the short interval between its completion and failure. The village of Desná (Dessendorf) is represented by (2),
and the Souš dam is represented by (3). The Bílá Desná reservoir was connected to the Souš reservoir by an underground tunnel. Source: Archive in the Jablonec nad Nisou.
137
3.2. Methods
3.2.1. Peak discharge estimation
Costa (1985) presented an overview of simple empirical methods
for estimating the peak discharge from man-made dam failures. We
have chosen only those methods to estimate the approximate peak discharge
at the Desná dam failure, for which all the data required for the
calculation are available. Methods that provided more conservative
(lower) peak discharges were chosen, namely, the methods presented
by Brater and King (1976), Kirkpatrick (1977), Price et al. (1977), the
Soil Conservation Service (1981), Wetmore and Fread (1981), and
their variations. For the dam height-based calculation (Kirkpatrick,
1977; Soil Conservation Service, 1981; Costa, 1985), the real dam height
(14.16 m) was also reduced by the dam freeboard (2.90 m) to 11.26 m
for the purpose of more conservative and representative results because
the reservoir was not full when the dam failed; thus, calculated results
of dam height-based calculations would be overvalued.
In addition to the empirical equation-based calculations, the
Hjulström curve, which describes the relation between the flow velocity
and the dimensions of transported boulders (Hjulström, 1935), and
graphical hydrograph-based estimations were used. The obtained results
were compared with recorded transversal profiles documenting
the maximal achieved water level during the flood. Hydrological significance
of the flood was investigated by confronting with the mean flow
rate and with the measured 1913 hydrometeorological flood peak discharge
of the Bílá Desná River. Classification of the geomorphic impacts
of the flood based on defined qualitative geomorphic alterations according
to Costa and O'Connor (1995) was used for selected parts of the affected
river channel.
3.2.2. Detection of erosion and of accumulation processes and modelling the
flood wave magnitude
Dominant geomorphic processes (erosion or accumulation) were
primarily recognised from photographs taken during the first few
months after the event and from technical maps created during the investigation
of the dam failure causes. Although the basic geomorphic
impacts related to erosion and accumulation processes have been identified
for the entire river reach, the detailed analysis of channel changes
and flood wave effects on the mid-segment of the river have been
Table 2
Basic characteristics of former Desná dam (according to Smrček, 1917; Žák et al., 2006).
Characteristic
Dam type Embankment (earthen) dam
Built 1911–1915
Dam height (vertical distance between
dam abutment and dam crest)
14.16 m
Dam crest width 172.8 m
Width of dam crest 5.2 m
Width of dam abutment 54 m
Distal face inclination 1:1.5
Proximal face inclination 1:1.5–1:2
Dam crest elevation 820.50 m asl
Elevation of maximal water level 818.90 m asl
Elevation of the bottom 806.34 m asl
Volume of dam body 31,920 m3
Maximal volume of accumulated water 400,000 m3
Catchment area 8.2 km2
Maximal discharge through the outlet 5.20 m3
s−1
Maximal discharge through the tunnel 36 m3
s−1
Maximal discharge through the spillway 37.6 m3
s−1
Table 3
Documentary data analysed within this study.
Origina
Title Year(s) Typeb
T-Dc
Source
Offic. Court reports (prosecution, defense) 1916–1932 man (Q)St Archive of the Jablonec nad Nisou (N Czechia)
Offic. Court reports by invited experts 1916–1932 man (Q)St Archive of the Jablonec nad Nisou (N Czechia)
Offic. Insurance investigation reports 1916–1917 man (Q)St Archive of the Jablonec nad Nisou (N Czechia)
Offic. Cadastral maps (1:2880) 1843 cart (Q)St Czech Office of Land Survey and Cadastre
Offic. 3rd Military maps (1:25,000) 1877–1880 cart (Q)St Österreichisches Staatsarchiv—Kriegsarchiv Vienna
Offic. Official photos of impacts icon (Q)St Archive of the Jablonec nad Nisou (N Czechia)
Offic. Aearial photos 1952, 2010 icon (Q)St Military Geographical and hydro-meteorological
Office of the Czech Republic
CSc. Prager Tagblatt (Katastrophalen Talsperren-Bruch) 1916 news (Q)C Web repository
CSc. Deutsche Bauzeitung (Der Talsperren-Dammbruch an der Weißen
Desse im Isergebirge am 18. September 1916)
1916 news (Q)C Librarian web repository (http://www.kobv.de)
CSc. Reichenberger Tageblatt (Zum Dammbruch der Weisen Desse-Talsperre) 1916 news (Q)C Web repository
CSc. Věstník organizačního komitétu Strany katolického lidu pro diecézi
královéhradeckou (Bulletin of the organizing committee of Catholics in
Hradec Králové diocese)
1916 news (Q)C Librarian web repository
(http://kramerius.lib.cas.cz/search)
CSc. Immergrün Illustrierte Kriegs-Chronik (Die Dammbruch-Katastrophe
in Dessendorf)
1916 pop (Q)C Archive of the Jablonec nad Nisou (N Czechia)
CSc. Der Talsperrenbruch an der Weissen Desse 1916 pop (Q)St Scientific library in Liberec (N Czechia)
CSc. Smrček A: O bezpečnosti zemních hrází (On safety of earthen dams) 1917 sci (Q)St Scientific library in Liberec (N Czechia)
CSc. Neustück Darre a.d. Talsperre 1920 pop (Q)St Scientific library in Liberec (N Czechia)
CSc. Gebauer E: Zum Dammdurchbruch der Talsperre an der Weissen Desse 1920 sci (Q)St Scientific library in Liberec (N Czechia)
CSc. Gnendiger E: Der Dammbruch der Talsperre an der Weissen Desse 1920(?) sci (Q)St Scientific library in Liberec (N Czechia)
CSc. Karpe L: Der Dammbruch an der Weissen Desse am 18. September 1916 1935 sci (Q)St Scientific library in Liberec (N Czechia)
CSc. Mikolášek V: Protržení přehrady na Bílé Desné 18. září 1916
(Dam-failure at the Bílá Desná River on 18th September 1916)
1966 pop (Q)St Scientific library in Liberec (N Czechia)
CSc. Žák L: Katastrofa na Bílé Desné—80. let
(Catastrophe on the Bílá Desná River—80 years)
1996 pop (Q)St Scientific library in Liberec (N Czechia)
CSc. Žák L et al.: Jizerskohorské přehrady a katastrofa na Bílé Desné—Protržená
přehrada (Dams in the Jizerské hory Mts. and the catastrophe on the
Bílá Desná River—dam failure)
2006 pop (Q)St Scientific library in Liberec (N Czechia)
CSc. Catastrophe at the Bílá Desná River—movie 1916 av (Q)St Web repository
P Photographic collection 1916 icon (Q)St Archive of the Jablonec nad Nisou (N Czechia)
P Postcards 1916 icon (Q)St Antiquarian web repositories
a
Origin = purpose of origin: Offic.—official, CSc—commercial and scientific, P—personal.
b
Type = data type: man—manuscript, cart—cartographic document, icon—iconographic document, av—audiovisual recording, sci—proto-scientific or scientific report in periodical or
monograph, news—newspaper report, pop—popular periodicals and books (regional literature).
c
T–D = time dimension of the source document: (Q)St—quasi-stationary and stationary data, (Q)C—quasi-continual and continual data (typology according to Raška et al., 2014).
138
performed in two detailed case studies. First, the channel changes after
the dam failure directly above and below the former dam have been
identified by a comparison of technical maps from 1916 and of aerial
photos from 1954 and 2010 in a GIS environment and corrected using
historical photos from archives. Second, the flood wave extent and its
effect on the mid-segment of the Bílá Desná River have been modelled
in GIS from a DEM (digital elevation model). Clearly, the DEM precision
is crucial for the flood wave modelling. To produce the most precise results,
we used a combination of a TIN (triangular irregular network)
model and a raster model (source data: the Czech Office of Land Survey
and Cadastre 1:10,000), which was derived using the spline method
with tension interpolation from original elevation data with a calculated
root mean square error (RMSE) of 1.10, i.e., adequate to the limit value
of 1.0 m sensu Svobodová (2008). We used the current digital elevation
data as no data are available for the years around the 1916 dam-failure
event. The scale of the data set was chosen to exclude possible influence
of channel modifications after the 1916 event (such as steps). The old
maps and aerial photos from the 1950s and the present day were compared
in order to identify the locations of anthropogenic transformations
of the surface (new buildings, roads, etc.), which may influence
the results of flood wave modelling. Finally, the water level during the
flood was derived from six preserved investigation reports prepared
for the insurance companies and included the detailed plans and
sketches of the buildings with water level marks. The water level data
were transformed to cross section profiles across the river valley, and
the water surface during the flood was created. The final flood wave extent
was modelled as a surface difference between the original earth
surface and the water surface and compared with current flood plans
of Q5 and Q100 floods (VÚV, 2014) to verify the correctness of the results
from the peak discharge estimation.
4. Results
The results of our investigation are divided into two subsections. The
first part is devoted to the reconstruction of the flood and its effects, exclusively
based on available documentary data sources (see Section 4.1).
The second part is focused on a supplementary description of the channel
evolution after the flood event, as documented by field survey and
documentary data (see Section 4.2).
4.1. Reconstruction of the flood and its geomorphologic and
landscape effects
4.1.1. Chain of events
The Bílá Desná dam construction began in 1911 because of discussion
concerning the construction of dams in the Jizerské Mts., which started
after catastrophic hydrometeorological floods at the end of June 1897
and finished four years later with dam approval on 18 November 1915
(Fig. 3). The longevity of the dam was only 10 months, before the dam
failed on 18 September 1916.
The chain of events describing the process of dam failure on 18
September 1916 in detail was outlined from existing data in the
literature (Smrček, 1917; Žák, 1996; Žák et al., 2006) and from archival
documentary data. The first evidence of piping through the dam was observed
by loggers returning from work along the dam at 15:30 h. At that
moment, 290,000 m3
of water was retained in the reservoir. The dam
keeper ordered the immediate draining of the reservoir by the full opening
of the dam outlet; nevertheless, the piping continued, and the flow
rate was exponentially increasing. At 16:00 h, the inhabitants living
downstream in the Desné village were warned for the first time by telegraph
against the increasing discharge from the dam. Fifteen minutes
later, at 16:15 h, the dam failed with an accumulated water volume of
250,000 m3
. The village of Desná inhabitants were warned again against
a flood following the dam failure at that time. An additional thirty minutes
later, at 16:45 h, the dam was completely empty (Fig. 4). The escaped
water swept through the valley, leaving behind 62 fatalities and
tremendous material damage.
Another point of view concerning the Desná dam failure is represented
by a hydrograph describing the change in the flow rate and the
decreasing volume of accumulated water between 15:30 h and 16:45
h on 18 September 1916 (Fig. 5). Costa (1985), and later by Tsakiris
and Spiliotis (2013), showed that hydrographs describing earthen
dam failures are particularly characterised by a steep rising limb and
by a steep falling limb, with a peak discharge within minutes to hours,
depending on the accumulated water volume, dam characteristics,
and failure mechanism.
4.1.2. Peak discharge
It is quite desirable to at least estimate the peak discharge, which is
not known exactly, and to put the peak discharge into the context of the
recorded geomorphologic and landscape effects of the event. The total
volume of 250,000 m3
was accumulated in the reservoir at the moment
of dam rupture and that the reservoir was empty in 30 min (Žák et al.,
2006). These numbers provide the mean flow rate of 138.9 m3
s−1
;
however, the peak discharge, which is responsible for most of the damage,
had to be several times greater. The rough graphical hydrographbased
estimation of the peak discharge (Fig. 5) showed that the peak
discharge apparently exceeded 1000 m3
s−1
, with a discharge of
57,000 m3
in 60 s (mean flow rate of 950 m3
s−1
). In addition, several
empirically based calculations (see Table 4) were used to approximate
the peak discharge. The obtained results ranged from 524.3 m3
s−1
to
1491.7 m3
s−1
at the dam. More conservative results, which were calculated
with a reduced dam height, ranged from 418.2 to 971.8 m3
s−1
.
Historical photographs of the flood effects, which were provided by
Gnendiger (1920), showed that the dimensions of the largest boulders
transported during the flood were greater than 2 m, with weights of
several tons. According to the Hjulström diagram (Hjulström, 1935),
to erode and to transport boulders of these given dimensions, the
mean flow velocity had to be greater than 4 m s−1
. Based on the historical
scheme of the failed dam, the breach area (the cross-sectional area
of gap in the dam measured from its abutment to the water level at the
time of the dam failure) was estimated to be 163.9 m2
. Clearly, the flow
rate may be calculated as a product of the breach cross-sectional area
and the flow velocity. These input data provide a peak discharge
Fig. 3. Chain of events connected to the Desná dam failure (according to Smrček, 1917; Žák et al., 2006).
139
N655.6 m3
s−1
, according to the mean flow velocity. This result corresponds
with the results of the more conservative numerical calculation
of the peak discharge (Table 4).
The flood wave following the dam rupture uprooted fully grown
trees growing downstream in the affected valley and transported
suspended material, including large boulders, with volumes up to
10 m3
. Smrček (1917) noted that transported wood, tree trunks and
large boulders dragged away by the flood wave caused repeated formations
and subsequent failures of temporary blockages in the upper part
of the mostly deep and narrow Bílá Desná valley. According to
Gnendiger (1920) in Žák (1996), eyewitnesses described these barriers
as at least 20 m high, and the sudden one-off escape of the temporarily
accumulated water most likely significantly increased the peak discharge
and might have caused several local peak discharges.
The volume of transported material considerably increased when
the flood wave struck the wood-processing factory (sawmill) situated
upstream of the village of Desná. Between 4000 and 5000 m3
of stored
wood was carried away (Mikolášek, 1966). We suppose that the
amount of dragged material and the effect of temporal barriers are
two of the most important reasons why the peak discharge was higher
than generally expected, particularly farther downvalley from the position
of the dam, and together are the reason why the flood had such catastrophic
effects on the village of Desná. A sudden increase in the flow
rate (rapid increase in the Jizera River water level for ~0.2 m) was also
registered in Mladá Boleslav, which is situated more than 60 km away
(Mikolášek, 1966).
To realise the hydrological significance and the magnitude of the
peak discharge achieved during the event, the obtained results should
be interpreted in the context of the mean flow rate of the Bílá Desná
River, which is 0.49 m3
s−1
, according to Vlček (1984). Therefore, the
most conservative calculated peak discharge (418.2 m3
s−1
; see
Table 4) exceeded the mean flow rate by more than 850 times. The
peak discharge of a ‘classical’ hydrometeorological flood that occurred
on 17 September 1913 was measured at 43 m3
s−1
(Smrček, 1917).
The most conservative calculated result overcame this hydrometeorological
flood peak discharge by almost 10 times.
4.1.3. Erosion and accumulation processes
The identification of erosion and accumulation processes, as well as
the effects of the flood on the infrastructure and buildings, was based on
a detailed analysis of the documentary data, which were primarily old
photos. Although the amount of available data is high (see Table 3,
Section 3.1), most of the photos depict the flood effects in detail
(e.g., an individual building from a close distance) and can be exactly located
with difficulty. This difficulty is also because of the absence of a
detailed map from the year of the flood event, which would provide information
regarding the spatial context of the depicted object. The older
cadastral maps (1840s) and, in turn, the later aerial photos (1950s) that
Fig. 4. Upstream view of the Desná dam before (A) and after (B) the dam failure.
Source: Archive of the Jablonec nad Nisou, Court reports.
Fig. 5. Approximate hydrograph of the event (based on the recorded time sequence of the dam failure).
140
are available did not correspond to the exact situation in 1916. Thus, the
combination of photos, different maps and audio–visual recordings, together
with written reports, were employed, and only the objects that
could be exactly localised were included in the evaluation of geomorphologic
and landscape effects.
First, the results of the overall assessment of flood effects along the
Bílá Desná River are presented, and then we focus on two detailed
case studies.
Fig. 6 shows the identified effects of the flood along the river between
the failed dam and the town of Tanvald. The detected effects include
(i) damage to buildings and to infrastructure and (ii) modifications of
the river channel (erosion and accumulation processes), and their intensity
can be divided into three segments relative to the longitudinal profile
of the Bílá Desná River (Fig. 1D). More than one-third of the
studied river length (0–3.2 km) is situated in a narrow intra-mountain
valley, with an average slope inclination of ca. 2.80° (806–650 m asl).
Table 4
Peak discharge estimation based on empirical calculations.
Author(s) of the method Equation Variable(s)
[units]
Value Calculated peak
discharge
[m3
· s−1
]
Soil Conservation Service (1981) (conservative
variation according to Costa (1985))
Qmax = 10.5 · H1.87
Qmax—peak discharge [m3
· s−1
]
H—dam height [m]
H = 11.26 m a
971.8
H = 14.16 m 1491.7
Kirkpatrick (1977) Qma x = 2.297 · (H + 1)2.5
Qmax—peak discharge [ft3
· s−1
]
H—dam height [ft]
H = 36.94 ft a
576.8
H = 46.46 ft 1009.3
Costa (1985) Qmax = 961 · V0.48
V—reservoir volume [106
m3
] V = 0.29 · 106
m3
530.5
Costa (1985) Qmax = 325 · (H · V)0,42
Qmax—peak discharge [m3
· s−1
]
H—dam height [m]
V—reservoir volume [in 106
m3
]
H = 11.26 m a
V = 0.29 · 106
m3
534.2
H = 14.16 m
V = 0.29 · 106
m3
588.2
Price et al. (1977) Qmax = 8/27 g1/2
· y3/2
· (0.4 b + 0.6 T) Qmax—peak discharge [m3
· s−1
]
g—gravitational acceleration [m · s−2
]
y—reservoir depth upstream [m]
b—width of breach base [m]
T—top width of breach [m]
g = 9.81 m · s−2
y = 10.26 m
b = 8 m
T = 30 m
646.6
Brater and King (1976);
Wetmore and Fread (1981)
Qmax = 3.5 · W · [c/(t + c/H0.5
)]3
, where c = 23.4 SA/W
Qmax—peak discharge [ft3
· s−1
]
W—average breach width [ft]
t—time of breach formation [hours]
H—dam height [ft]
SA—reservoir surface area [ac]
W = 62.3 ft
t = 0.75 h
H = 36.94 ft a
SA = 24.71 ac
418.2
W = 62.3 ft
t = 0.75 h
H = 46.46 ft
SA = 24.71 ac
524.3
a
Reduced value (see Methods section).
Fig. 6. Effects of the flood along the Bílá Desná River based on documentary data. Note: Of the large amount of documentary data, only those data that could be exactly localised were used.
141
The high gradient of this river segment enabled the stream to transport a
large amount of material, which consisted of fractions varying from sand
to boulders of up to 2 m in diametre. The flood wave damaged the forest
ecosystem in the narrow valley, and the channel underwent a lateral
shift (see below in case study 1) and a slight incision in some segments.
After reaching the end of the narrow valley, the Bílá Desná River reaches
the open space with the settlement of the village of Desná. Between 3.2
and 5 km, the gradient slightly increases to ca. 3.8° as the river channel
enters the foothills; however, this gradient gradually decreases from
5 km downstream. The decrease in the gradient, which causes a loss of
flow velocity, resulted in the gradual, but massive, deposition of boulders
and of timber (approximately between B and D in Fig. 6) that damaged
or destroyed tens of buildings, including a few industrial facilities
(Gnendiger, 1920) in this river segment. The woody buildings localised
near the channel and in the open floodplain (D in Fig. 6) were primarily
affected. Owing to the transformation of the flood wave, the last analysed
segment of the river (approximately within case study 2) was primarily
affected by flooding; however, the destruction of buildings was not frequently
documented. This lack of documentation can be ascribed to not
only the increasing share of buildings constructed from stones and bricks
in this segment but also the decreasing volume of boulders and timber
that reached the lower parts of the river. In contrast, the documentary
evidence has shown that part of this transported material (both timber
and boulders) reached as far as the railway bridge in Tanvald (E in
Fig. 6), ca. 8 km below the failed dam.
4.1.3.1. Case study 1. The disturbance and modification of the valley and
riverbed of the Bílá Desná River following the dam failure has been reconstructed
in case study 1, which covers the area of the former reservoir
(see Fig. 1). The detailed reconstruction of other river reaches could not
be performed because of the limited availability of old maps and photos.
Following the dam failure and the rapid discharge of 250,000 m3
of
accumulated water during only 30 min, the channel below the dam
was significantly modified. The comparison of historical maps from investigation
reports of the event and of aerial photos from the early
1950s and from 2010 has shown that the meandering channel below
the dam was subject to channel migration and lateral accretion (Fig. 7).
The channel width in the reach below the former reservoir varied around
5 m before the dam failure and is currently 8 m in average. The magnitude
of channel migration reached up to 30 m in the second meander
below the former dam, whereas the lateral accretion only slightly increased
the sinuosity of the channel. Instead of channel straightening
(cf. Pizzuto, 2002), which could be expected as a result of flood wave
flow, the return to the original meandering channel seems to be caused
by geometry of a dam-breach and by extreme discharge that disrupted
the channel geometry artificially established during the dam construction.
This seems to be caused by the disrupting impact of the flood
wave caused to channel segments. As documented by cadastral maps
from the mid-nineteenth century, the Bílá Desná River channel approximately
returned to its original position before the dam construction,
with a meandering pattern. The reestablishment of the meandering
Fig. 7. Changes in the channel morphology after the dam failure.
142
channel pattern and increased sinuosity by lateral accretion enabled
the evolution of several gravelly points and irregular bars, which are
characteristic of other high-gradient streams in the area. According to
Costa and O'Connor (1995), the above-described geomorphic changes
are characterised as ‘extreme disruptions’ of the river channel.
4.1.3.2. Case study 2. The flood wave extent was reconstructed to estimate
the effects of the dam failure on towns downstream of the Bílá
Desná River and to verify the reliability of the peak discharge calculated
from empirically based equations and from other methods used (4.1.2).
Because of the incompleteness of documentary data that would document
flood effects along the entire Bílá Desná River, the reconstruction
was only performed for a selected river reach between the towns of
Potočná (southern margin of the village of Desná) and Tanvald (case
study 2; Fig. 1). Six preserved investigation reports for insurance companies,
which included detailed sketches and cross profiles of the
flooded buildings, enabled the reconstruction of eight cross profiles indicating
the water level. Using these profiles, the flood wave extent
was estimated for the river reach of ~2 km in length (Fig. 8). All of the
flooded buildings used for the flood wave estimation are still standing
and the surface around them has not been modified significantly. By
comparison of old and present maps and aerial photos, we identified
only three localities within the estimated 1916 flood wave extent that
were subject to further anthropogenic transformation of the surface
(white dots in Fig. 8). The modelled flood wave estimation can therefore
be considered relevant, despite the absence of historical elevation data.
The width of the flooded area ranged between ca. 50 and 250 m,
with the highest width reached in built-up areas of the towns and in
the meandering channel in the central part of the studied channel
reach. In contrast, the minimal width was typical of narrow valley
segments, where an increase in the flow velocity can be expected. However,
the documentary data did not provide the cross profiles for the
central part of the studied river reach. Therefore, the results for this
part of the area must be considered only rough estimations. According
to the classification presented by Costa and O'Connor (1995), this part
of the Bílá Desná River channel was affected by ‘small disruptions’
from the point of view of geomorphic impacts.
The comparison of the reconstructed flood wave extent with available
current flood maps (VÚV, 2014) has shown that the flood wave following
the 1916 dam failure exceeded the width of Q5 and Q100 floods
10 times and 5 times, respectively, at some locations. This result can
be ascribed to both the extremeness of the peak discharge reached
after the dam failure and the subsequent modifications of the Bílá
Desná River channel during the last 100 years.
4.2. Contemporary field evidence and later modifications of the
river channel
4.2.1. Artificial river channel modifications
After the catastrophic flood in 1916, the primarily affected section of
the Bílá Desná River channel (from the upper part of the village of Desná
to the confluence with the Černá Desná River; see Fig. 6) was artificially
Fig. 8. Estimated flood wave extent. Part (A) illustrates the flood wave extent and the Bílá Desná River valley cross profiles. White dots for major changes in built-up areas after 1916 within
the estimated flood wave extent (see Section 4.1.3 for further explanation). Examples of documentary evidence for the flood wave extent and the water level, as preserved in investigation
reports for insurance companies. Cross-profile (B), plan (C), and photo (D) with an indicated water level mark.
143
modified between 1923 and 1936. These technically demanding and expensive
remedial works included the complete pavement of the river
bottom along the 1.5-km-long lower part and the construction of 34
stony steps in the steeper 2-km-long upper part (Fig. 9A); bedrockbased
parts of the river bottom were left without modifications
(Fig. 9B). The stony steps are between 1.0 and 4.2 m high and 11.0 to
18.0 m long and serve as a sediment trap (prevention against the
initialisation of debris-flow movement). All of these artificial elements
are designed for a maximal flow rate of 60 m3
s−1
(Žák, 1996). The section
from the upper part of the village of Desná to the former Desná dam
position has a natural character, with no artificial channel modifications
(Fig. 9C).
4.2.2. Effects of later hydrometeorological floods on the river channel
During the decades after the 1916 event, the Bílá Desná River has
been affected by several flash flood events summarised in Table 5.
Only during the last three decades (since the early 1980s) has the maximum
daily rainfall exceeded 100 mm in seven years (Kulasová et al.,
2006). Among the most serious events that have had significant effects
on the river channel and its surroundings, the floods in 2002 and 2010
must be mentioned. The summer 2002 flood struck almost the entire
territory of Czechia, with significant socioeconomic impacts on small
mountain watersheds and on the largest rivers, such as Vltava and
Labe. The daily maximal rainfall in the Jizerské hory Mts. was 278 mm
(Kulasová et al., 2006), and some of the artificial steps at the Bílá
Desná River channel were damaged during the subsequent flood and
needed to be repaired (Žák et al., 2006). In 2010, two major flash
flood events occurred in the territory of Czechia (Daňhelka and Šercl,
2011). Although the first flash flood following the heavy rainfall in
May 2010 primarily affected the eastern part of the country (Moravia),
the flash flood in August 2010 was territorially concentrated in the
northern part of the country, with a daily rainfall maximum of
179 mm (Kašpar et al., 2013). The severity of this unanticipated event
was also characterised by its 100-year rainfall recurrence interval, in
contrast to the Moravian floods in May with a recurrence interval of
20 years. The flash flood event again damaged or destroyed artificial
modifications of the Bílá Desná River channel and the channel surroundings
(Collective, 2010).
Considering these flash flood and flood events during the last century,
therefore, clearly field evidence for the 1916 flood cannot be easily
identified in the current terrain. This difficulty in identification holds
true not only in the inhabited segments of the Bílá Desná River valley
(the towns of Desná and Tanvald), with significant anthropogenic
Fig. 9. Part (A) shows an example of stony steps built into the steeper part of the Bílá Desná River channel to protect against floods (debris flows). Part (B) shows a natural bedrock-based
part of the river channel bottom. Part (C) shows a natural (unremediated) part of the Bílá Desná River channel.
Table 5
Extreme rainfalls (more than 150 mm/d) in the Jizerské hory Mts. between
1916 and 2014 followed by flood and flash flood events; compiled from
Štekl et al. (2001), Bubeníčková et al. (2003), Kulasová et al. (2006), and
Kašpar et al. (2013).
Year Maximal daily rainfall totals [mm]
1920 214.5
1934 151.0
1937 169.0
1958 153.2
1992 154.9
1997 162.8
2002 278.0
2010 179.0
144
transformation and with the increase in built-up areas during the twentieth
century, but also in noninhabited segments of the river valley. In
the noninhabited river valley segments, the anthropogenic modification
of the channel is accompanied by a succession in permanent grasslands
and in mixed-forest areas surrounding the channel and affected by the
flood events.
5. Discussion
5.1. Reliability and completeness of documentary data
Documentary data, including both iconographic and written sources,
have been widely employed in the research of different types of natural
hazards (Hooke and Kain, 1982; Glade et al., 2001), such as meteorological
extremes and their relation to climate change (e.g., Pfister et al.,
2008; Brázdil, 2009), historical floods (e.g., Guzzetti et al., 1994;
Brázdil and Kundzewicz, 2006), earthquakes (e.g., Cammasi, 2004) or
landslides (e.g., Ibsen and Brunsden, 1996; Raška et al., 2014). Major attention
has been devoted to continual (periodical) data that enable
parametric time-series to be built as a basis for magnitude–frequency
analysis. The reconstructions of individual natural hazard events necessitate
the integration and critical analysis of various data types (continual
and stationary) and raise new questions regarding data availability
and reliability. With respect to data availability, this study shows that
a broad spectrum of sources exists. This data availability can be ascribed
to the fact that the causes of the catastrophic flood at the Bílá Desná
River were originally assigned to human fault and resulted in the systematic
collection of documentary evidence and reports by the court.
This situation is in contrast with historical hazards with natural causes
in Czechia, which are often documented by photos and by descriptions;
however, these historical hazards were scarcely subject to systematic
study and interpretation.
A significant share of the available sources used in this study was
only accessible via personal detailed searches in regional archives. The
digital on-line catalogues (if any) contain only brief descriptions of socalled
funds, i.e., parts of archival sources devoted to a certain topic
(e.g., local administration in 1848–1918) that include hundreds of
unsorted documents. Moreover, the study illustrates, that although
the amount of documentary data may be sufficient to provide the
basic information concerning the individual catastrophic event into
the time-series, the data do not provide all information necessary to reconstruct
the causes, impacts, and recovery after the individual event.
Analogous to the higher frequency of available data in regions or periods
more prone to these events (e.g., Tropeano and Turconi, 2004), we have
documented differences in data availability within the areas affected by
an individual event.
When documenting flood impacts along the studied part of the Bílá
Desná River, distinct differences were found in the amount and quality
of preserved data for individual river segments. Those segments that
suffered from the most serious impacts were documented by many
photos and comments, which, in contrast, could not be frequently localised
accurately, and their paraphrasing in further texts resulted in informational
ambiguities. The downstream segments in the towns of Desná
and Tanvald, characteristic of factories, manufacturers, and houses of
the rich social class (built from bricks and stones), suffered from
minor effects in terms of the number of victims and the number of
destroyed buildings. Therefore, these segments were documented by
less sources but, simultaneously, with high precision, enabling the reconstruction
of the flood wave extent, for instance. These findings confirm
the fundamental sensitivity of documentary data to differences in
social vulnerability and in the perception of risks. Therefore, these
data must be understood as a social construct created by historical communities
and influenced by their perception of the event rather than as
parametric information suitable for interpretation on a statistical basis
(Stanford, 1986; Raška et al., 2014).
5.2. Accuracy of the hydrometrical estimations
Peak discharge is generally calculated as a passed volume of water divided
by a given time interval. Because the dam discharge following the
dam failure is a nonlinear process, the peak discharge is subordinated to
the exact definition of the time interval. The typical hydrograph curve
shape of the dam failure posed by Costa (1985) clearly shows that the estimated
peak discharge will increase together with a shortening of the
time interval used for the calculation (see also Fig. 4). Therefore, only
the hydrograph-based estimation of the peak discharge is quite susceptible
to error.
The empirical equation-based calculation of the peak discharge
based on different input data [volume-based equation (Costa, 1985),
the dam height-based equation (Kirkpatrick, 1977; Soil Conservation
Service, 1981), the breach shape-based equation (Price et al., 1977),
and a time-dependent equation (Brater and King, 1976; Wetmore and
Fread, 1981)] were also used to approximate the peak discharge. Nevertheless,
all these methods are subordinated to the uniformity and representativeness
of the original data sets used for their construction. The
last method used (Hjulström curve-based estimation; Hjulström,
1935) is particularly limited for extreme flow velocities and/or for extreme
grain sizes.
As shown above, each type of method used for peak discharge estimation
has its own potential sources of errors. However, we suppose
that the combination of all of the above-mentioned methods limits
the error rate of each method and provides useful results (proven reliability
of mutually obtained results).
5.3. Comparative perspective on the event magnitude and impacts
In comparison with recorded floods following dam failures in
Europe, the Bílá Desná dam failure claimed a high number of fatalities
relative to the volume of escaped water (Fig. 10). The discharged volume
of 290,000 m3
and 62 fatalities provide an extremely catastrophic
ratio of 4677 m3
/victim. From this point of view, the Bílá Desná dam failure
is considered one of the worst dam failures ever recorded worldwide.
In the regional context of the Czech Republic, no recorded
hydrometeorological flood (e.g., the great floods in 1997 and in 2002)
claimed such fatalities. According to the number of fatalities, this
event is currently ranked among the most catastrophic earthen dam
failures in Europe. When we consider the time of the dam failure, this
event was ranked as one of the most catastrophic dam failures until
1916 and, therefore, significantly affected the engineering of earthen
dams in the entire Central Europe region (Žák, 1996).
From the perspective of the achieved peak discharge (or the ratio of
the peak discharge to the mean flow rate), the flood following the Bílá
Desná dam failure is characterised by extreme values, which are out of
the order of magnitude of ‘classical’ hydrometeorological floods in the
context of the hydrological regime of rivers in the Central Europe region.
Fig. 10. Comparison of the most catastrophic European dam failures. (Data: Harrison,
1974; Costa, 1985; Alcrudo and Mulet, 2004; Peng and Zhang, 2012).
145
These findings correspond with those findings presented, e.g., by Costa
(1985), who noted that this characteristic is a typical feature of floods
following man-made dam failures. Our investigation also showed that
from the point of view of geomorphic impacts, the Bílá Desná dam failure
is characterised by rather expectable changes in the river channel (valley)
in comparison with comparable recorded dam failures worldwide
and with regard to the event magnitude (e.g., Costa and O'Connor,
1995; Bathurst and Ashiq, 1998; The Dolgarrog Disaster, 1925).
6. Conclusions
In addition to the often catastrophic social and economic effects,
floods following man-made dam failures are also characterised by significant
geomorphic and landscape effects, which are frequently
disregarded. Our investigation of the catastrophic flood effects following
the Bílá Desná dam failure on 18 September 1916 showed the great potential
of documentary data sources for the reconstruction of an individual
catastrophic event, where the direct geomorphic and landscape
evidence has been covered by human activities. First, the analysis of
these data enabled the estimation of the peak discharge during the
flood, the documentation of the geomorphic effects of the flood on the
channel morphology downstream of the dam and the reconstruction
of the flood wave extent in downstream segments of the Bílá Desná
River. Considering the acquired results, the flood following the Bílá
Desná dam failure may rank among the most serious disasters in
terms of the number of victims per discharged water volume that
have occurred worldwide during the twentieth century. These results
also provide insight into the vulnerability of historical communities to
natural hazards and into their recovery measures. Second, the performed
research confirmed previous studies emphasising the critical
analysis of documentary data as a necessary step for their employment
in geoscientific studies.
Acknowledgement
References
Alcrudo, F.,Mulet, J., 2004. The Tous dam break case study. IMPACT project—Investigation
on extreme Flood Processes and Uncertainty.
Bathurst, J.C.,Ashiq, M., 1998. Dambreak flood impact on mountain stream bedload transport
after 13 years. Earth Surf. Process. Landf. 23 (7), 643–649.
Brater, E.F.,King, H.W., 1976. Handbook of Hydraulics. McGraw-Hill Book Co., New York.
Brázdil, R., 2009. Historical climatology and its contribution to climate reconstruction
in Europe. In: Kirchhofer, A. (Ed.), Nachhaltige Geschichte. Chronos Verlag, Zurich,
pp. 65–89.
Brázdil, R., Kundzewicz, Z.W., 2006. Historical hydrology—editorial. Hydrol. Sci. J. 55,
733–738.
Bubeníčková, L., et al., 2003. Flood precipitation analysis in the Jizerské Mountains. In: NE
FRIEND Project 5, ERB. Proceedings of Workshop on Mountain Hydrology, Bucharest,
p. 26.
Cammasi, R., 2004. Catalogues of historical earthquakes in Italy. Ann. Geophys. 47,
645–657.
ČHMÚ, 2013. Mean annual precipitation in the Czech Republic (1961–2000). Available
online from http://portal.chmi.cz/files/portal/docs/meteo/ok/images/sra6190.gif
(accessed: 13.12.2013).
Collective, 2010. Summary report on flood on 7th–9th of August 2010 at the Labe and
Ohře Rivers. Český hydrometeorologický úřad, Hradec Králové, (in Czech).
Costa, J.E., 1985. Floods From Dam Failures. United States department of the interior geological
survey, Denver (Colorado).
Costa, J.E.,O'Connor, J.E., 1995. Geomorphically effective floods. In: Costa, J.E., et al. (Eds.),
Natural and Anthropogenic Influences in Fluvial Geomorphology. U.S. Geological
Survey, Vancouver, Washington, pp. 45–56.
Daňhelka, J., Šercl, D., 2011. Floods in the Czech Republic in 2010. Meteorol. Zpr.—
Meteorol. Bull. 64 (1), 4–9.
Gebauer, E., 1920. Zum Dammdurchbruch der Talsperre an der Weissen Desse. B. Förster,
Nieder Rochlitz.
Glade, T., Albini, P., Frances, F., 2001. An Introduction to the Use of Historical Data in
Natural Hazard Assessment. In: Glade, T., Albini, P., Frances, F. (Eds.), The use of historical
data in natural hazard assessments. Kluwer, Dordrecht, pp. xvii–xxv.
Gnendiger, E., 1920. Der Dammbruch der Talsperre and der Weißen Desse am 18.
September des Krieghjahres 1916Gebrüder Stiepel Ges, Reichenberg (Liberec).
Guzzetti, F., Cardinali, M.,Reichenbach, P., 1994. The AVI Project: a bibliographical and
archive inventory of landslides and floods in Italy. Environ. Manag. 18, 623–633.
Harrison, S., 1974. A Complete History of Great Flood at Sheffield on March 11 and 12,
1854. Evans & Longley Associates, Earlsheaton, Dewsbury, Yorkshire.
Hjulström, F., 1935. Studies of the morphological activity of rivers as illustrated by the
River Fyris. Bull. Geol. Inst. Univ. Upps. 25, 221–527.
Hooke, J.M.,Kain, R.J.P., 1982. Historical Change in the Physical Environment: A Guide to
Sources and Techniques. Butterworth Scientific, London.
Ibsen, M.L.,Brunsden, D., 1996. The nature, use and problems of historical archives for the
temporal occurrence of landslides, with specific reference to the south coast of
Britain, Ventnor, Isle of Wight. Geomorphology 15, 241–258.
Ingram, M.J.,Underhill, D.J.,Farmer, G., 1981. The use of documentary sources for the study
of pastclimates. In: Wigley, T.M.L., Ingram, M.J., Farmer, G. (Eds.), Climate and History.
Cambridge University Press, Cambridge, pp. 118–213.
Jarrett, R.D.,Costa, J.E., 1986. Hydrology, Geomorphology, and Dam-Break Modeling of the
July 15, 1982 Lawn Lake Dam and Cascade Lake Dam Failures, Larimer Country.
Colorado. U.S. Geological Survey Professional Paper 1369, Washington.
Karpe, L., 1935. Der Dammbruch an der “Weissen Desse”am 18. September 1916. H.
Dominicus Nachf, Teplitz-Schönau (Teplice).
Kašpar, M.,Müller, M.,Pecho, J., 2013. Comparison of meteorological conditions during
May and August 2010 floods in Central Europe. AUC Geogr. 48 (2), 27–34.
Kirkpatrick, G.W., 1977. Evaluation Guidelines for Spillway Adequacy. The evaluation of
DamSafety, Engineering Foundation Conference, Pacific Grove, California, Proceedings.
American Society of Civil Engineers, NewYork, pp. 395–414.
Kulasová, A.,Pobříslová, J.,Jirák, J.,Hancvencl, R.,Bubeníčková, L.,Bercha, Š., 2006. Experimental
hydrological base in the Jizera Mountains. J. Hydrol. Hydromech. 54 (2),
163–182.
Mikolášek, V., 1966. Protržení přehrady na Bílé Desné 18. září 1916 (Dam-failure at the
Bílá Desná River on September 18th 1916, in Czech). Klub pracujících Desná, Desná
v Jizerských horách.
Peng, M.,Zhang, L.M., 2012. Analysis of human risks due to dam-break floods—part 1: a
new model based on Bayesian networks. Nat. Hazards 64 (1), 903–933.
Pfister, Ch.,Luterbacher, J.,Wanner, H.,Wheeler, D.,Brázdil, R.,Ge, Q.,Hao, Z.,Moberg, A.,
Grab, S., del Prieto, M.R., 2008. Documentary evidence as climate proxies. White
paper written for the proxy uncertainty workshop in Trieste, 9–11 June 2008.
Pizzuto, J., 2002. Effects of dam removal on river form and process. Bioscience 52 (8),
683–691.
Price, J.T.,Lowe, G.W.,Garrison, J.M., 1977. Unsteady flowmodeling of dam-break waves.
Dam-break Flood routing Model Workshop Proceedings. National Technical Information
Service, Springfield, pp. 275–437.
Raška, P.,Zábranský, V.,Dubišar, J.,Kadlec, A.,Hrbáčová, A.,Strnad, T., 2014. Documentary
proxies and interdisciplinary research on historic geomorphologic hazards: a discussion
of the current state from a central European perspective. Nat. Hazards 70 (1),
705–732.
Smrček, A., 1917. O bezpečnosti zemních hrází pro vodní stavby a příčinách katastrofy na
Bílé Desné (On safety of earthen dams and causes of the Bílá Desná catastrophe, in
Czech). Česká matice technická, Praha.
Šobr, M.,Jeníček, M.,Vočadlová, K., 2008. Dam breach reconstruction of the Blažnov pond
caused by extreme hydrological event. AUC Geogr. 43, 105–114.
Soil Conservation Service, 1981. Simplified Dam-breach Routing Procedure. U.S. Department
of Agriculture, Washington D.C.
Stanford, M., 1986. The Nature of Historical Knowledge. Blackwell, Oxford.
Štekl, J.,Brázdil, R.,Kakos, V.,Jež, J.,Tolasz, R.,Sokol, Z., 2001. Extreme daily precipitation on
the territory of the Czech Republic in the period 1879–2000 and their synoptic causes
(in Czech). Národní klimatický program České republiky. 31, (Praha).
Svobodová, J., 2008. Hodnocení přesnosti digitálních modelů reliéfu (Assessment of the
DEM precision, in Czech). Geomorphol. Slovaca Bohem. 8, 76–81.
Tropeano, D.,Turconi, L., 2004. Using historical documents for landslide, debris flow and
stream flood prevention. Applications in Northern Italy. Nat. Hazards 31, 663–679.
Tsakiris, G.,Spiliotis, M., 2013. Dam-breach hydrograph modelling: an innovative semianalytical
approach. Water Resour. Manag. 27 (6), 1751–1762.
Vlček, V., 1984. Zeměpisný lexikon ČSR—vodní toky a nádrže (Geographical lexicon of
Czechoslovakia—Rivers and water reservoirs, in Czech). Academia, Praha.
VÚV, 2014. Flood discharge spatial models for the Czech Republic. GIS
datasetHydroecological information systemVýzkumný ústav vodohospodářský T.G.
M., Praha.
Wetmore, J.N.,Fread, D.L., 1981. The NWS simplified dam-break flood forecasting model.
Canadian Hydrotechnical Conference Proceedings.
Žák, L., 1996. Katastrofa na Bílé Desné—80 let (Catastrophe on the Bílá Desná River—
80 years, in Czech). Povodí Labe, a. s, Hradec Králové.
Žák, L.,Bláha, J.,Jaroušek, J.,Charvátová, E.,Melanová, M.,Rozsypal, A.,Simm, O.,Vinklát, P.D.,
2006. Jizerskohorské přehrady a katastrofa na Bílé Desné—protržená přehrada (Dams
in the Jizerské hory Mts. and the catastrophe on the Bílá Desná River—dam failure, in
Czech). KNIHY 555, Liberec.
146
Other sources
Archive of the Jablonec nad Nisou: investigation reports, judicial protocols and
documentation.
Chronicle transcript of municipality Nebřežiny. Available online [accessed: 13.1.2014]:
bhttp://www.nebreziny.cz/2005/povoden-r-1872-v-nebrezinech-podle-nebrezinske-
kronikyN.
Soběnov Municipality webpage. Available online [accessed: 13.1.2014]: bhttp://www.
sobenov.cz/misN.
The Dolgarrog Disaster—The website for Welsh cultural history. Available online
[accessed: 25.5.2014]: bhttp://education.gtj.org.uk/en/filmitems/29132N.
Tourist information portal turistika.cz. Available online [accessed: 13.1.2014]: bhttp://
www.turistika.cz/mista/strazny-prehradaN.
147